What the Engine does
It measures fame as a worldwide human reality, not as a social media contest.
A follower count can show popularity on one app. A chart can show what people are consuming in one industry. Neither can tell the whole world who is truly famous. The Global Fame Engine looks across many kinds of evidence, countries, languages, cultures, industries, and generations to create a much bigger and more accurate picture.
It is designed to understand living stars, breakout names, historical figures, public leaders, athletes, entertainers, creators, and people whose fame exists strongly in real life even when they have little or no personal online presence.
What iFANN means by fame
Fame is how widely the world knows who or what an entity is.
iFANN defines fame as the extent to which a distinct person, organization, brand, product, work, or other entity is recognized or recalled by the public across populations, places, languages, and time.
Recognition can mean knowing an entity by name, recognizing a person by face, recognizing a brand, product, or work through its visual identity or other defining cues, or having fuller knowledge of who or what the entity is. These are different forms and depths of public awareness.
The Engine estimates that recognition from independent evidence of lasting awareness, geographic and linguistic breadth, information-seeking, media visibility, and audience behavior. It does not claim that every signal is a direct recognition survey.
Attention is evidence of fame and can strengthen or weaken recognition, but attention is not fame itself. The Famous rankings measure sustained public recognition; Trending measures who is gaining attention fastest.
A worldwide view of fame
The Engine does not assume that the loudest corner of the internet represents Earth. It pays attention to fame across different countries and locations, from worldwide household names to people whose importance becomes clear through strong local and regional recognition.
This wider view helps iFANN publish worldwide, regional, country, and industry rankings while still recognizing that fame can cross borders, live deeply inside a culture, or endure across generations.
The trails it can understand
The Engine learns from many independent public data trails. Examples include Wikipedia, Google Trends, audience consumption, and global and local news: along with many other reliable sources and signals.
A broad view of what people around the world actively search for.
Current readership plus durable cross-language reference presence: evidence that recognition extends beyond repeated searches.
Evidence of recognition and attention across countries, cities, languages, and communities.
Real audience activity around music, film, television, video, sport, and other relevant fields.
These examples explain the range of evidence in ordinary language. The Engine's full data system and ranking logic are proprietary.
Built to be trusted
No person, platform, or single number gets to define fame on its own. The Engine checks the bigger picture, refreshes as the world changes, and is built to resist misleading spikes, narrow audiences, incomplete trails, and one-platform bias.
No one can buy a position. iFANN likes, followers, advertising, partnerships, and editorial preference do not decide the rankings.
More than a list
The Global Fame Rankings are the first major product powered by the Engine. The underlying technology is being built as a lasting data infrastructure that can eventually support research, media, licensed data products, and a public API.
The Global Fame Engine is the technology. A Global Fame Score is a number it produces. The Global Fame Rankings are the public lists people see.
Engine updates
The current release is Global Fame Engine v5.70. As the technology improves, this page automatically shows the newest version and a plain-language record of major releases.
Repair foreign evidence coverage and affinity floors
Uniform country worldwide support reduction
Restore country support balance
Undo the 90% home-country-only reduction from 5.67. Restore the preceding support weights across countries. Future country support multipliers are nationality-neutral; smaller reductions remain dry-run scenarios.
Local evidence leads home-country fame
Reduce both worldwide recognition-stock and broader-global support by 90% for people ranked in their own country, uniformly across all country Famous IN lists. Other scope formulas and foreign-person support are unchanged.
Cultural recognition first-place policy
Jesus receives editorial first-place priority in 26 audited country rankings covered by the existing cultural-affinity rule. Original computed scores and the editorial adjustment are retained separately. This priority reflects cultural judgment, not new measured evidence or a population-recognition percentage. Sri Lanka, other countries, Global, Famous FROM and Trending are unchanged by this policy.
Nigeria editorial recognition adjustments
Nigeria now includes explicit editorial adjustments of plus five index points for Muhammad and minus eighteen for Victor Osimhen. These judgments are recorded separately from source measurements, with the pre-adjustment score and rationale retained for audit. They are not new measured evidence or recognition percentages. Historical weighting, zero music-chart scoring, office neutrality and the corroborated foreign-evidence policy remain unchanged. Other countries, Global and Famous FROM are unaffected.
Nigeria office neutrality and corroborated local recognition
Holding national office no longer adds a special score in Nigeria. Foreign figures with durable Nigerian search and at least 24 readership days across two local sources retain at least 75% of their direct Nigerian evidence; global recognition and broader-global support keep their existing geographic discounts. Nigeria's 80% historical search weighting and zero music-chart score remain in force. These rules apply across professions, with no named-person score or rank assignment. Other countries retain their policies.
Nigeria historical recognition takes precedence
Famous IN Nigeria now assigns 80% of its search component to durable historical evidence and 20% to recent attention for every profession. Music charts no longer contribute score, source confidence, or chart-based qualification to that list. Their separate Trending role is preserved. Other countries, Global and Famous FROM retain their scoring rules. No individual score or rank is assigned manually.
Durable rankings resist short-lived news and recalibration
Famous is a measure of enduring recognition, so a burst of recent news coverage now moves an established name far less: the permitted news lift above an entity's own historical baseline was reduced to a third of its previous allowance, with the full undiscounted burst still powering the separate Trending product. The search-news historical baseline now spans the same year-long window as the article-volume lane rather than a rolling month, so a multi-week news cycle can no longer become its own baseline and bypass the cap. The long-run search-history foundation additionally gained a drift limiter, so a data-provider recalibration of the 2004-present series can shift an established entity's durable-search evidence only gradually across monthly collection cycles instead of reordering neighbors overnight.
Non-destructive source continuity
Failed or implausibly low weak-alias news refreshes no longer delete the last successful measurement. Mononym queries prefer identity-bearing aliases over unrelated nicknames, recent source-specific search-news evidence is recovered for up to ten days without refreshing its observation clock, and the publication gate rejects a flagship list that silently loses eligible recent news evidence behind another source.
Rap-primary continental cohort classification
Restricted the stronger rapper continental-recognition rule to rap-primary musicians. The first explicit singer, vocalist, or rapper role in recorded profession metadata determines the primary career, while reviewed singer/rapper hybrids may be exempted in transparent methodology configuration. Singer-first artists receive the ordinary living-musician rule; rap-first artists retain the stronger test. Drake is recorded as the reviewed singer/rapper hybrid exemption requested by the owner.
Worldwide recognition test for technology figures
Extended the Africa and Asia continental-recognition requirement to every technology-industry figure, living or deceased. Previously only deceased technology figures took the test, which left living technology and business leaders as the only people in the global top ranks whose worldwide recognition was never verified against the 20-country African and 27-country Asian search panels. The test measures each subject against their own positive-market reference, so genuine African and Asian recognition is rewarded and weak recognition is penalized up to a 7% continuous cap, identical to the policy already applied to deceased technology figures.
Country-isolated local ranking policy layer
Introduced configurable country-only ranking rules that apply exclusively to one country's Famous IN scope, never to Global, Famous FROM, regions, or other countries. Nigeria now blends its complete comparable search cohorts at 25% history and 75% recent recognition, retains the standard 80% historical durability for event-sensitive politicians, and exempts Jesus from the Nigeria origin-person leadership cap based on documented foundational cultural-religious salience. The resulting evidence-derived Nigeria leaders are Jesus, Davido, and Wizkid, with Davido and Wizkid separated by a deliberately narrow natural evidence margin.
Symmetric foundational-religion affinity and Nigeria cohort completion
Expanded local cultural-historical affinity for Jesus and Muhammad symmetrically using Pew Research Center's 2020 country religious-composition estimates: a figure receives full local-evidence treatment where the corresponding religion is the country's largest affiliation or has at least 10 million adherents. Completed Nigeria's current 90-day search comparison from 38 to 260 entities on one stable Burna Boy anchor, making current Nigerian search comparable again. Global scoring remains unchanged.
Restored unified Global and People ranking pool
Fully rolled back the v5.54-v5.55 split-score architecture. All and People again use the same penalty-adjusted Global Fame Score and the same Global Top 1,000 pool, exactly restoring the v5.53 ranking behavior. Consequently, People-specific Africa/Asia and intent adjustments also affect those people in All. The K-pop Wikipedia/search factor is restored to 80% for both people and groups.
Exact People and All policy isolation
Restored the People scoring path exactly to its v5.53 policy, including the 80% K-pop Wikipedia/search reliability factor, while retaining the separately calibrated 70% K-pop factor only in the mixed All leaderboard. Person-only and category-specific Africa/Asia and intent scrutiny remains exclusive to People. This corrects the v5.54 coupling that unintentionally changed K-pop People positions; non-K-pop People scores remain identical to v5.53.
Category-neutral mixed-entity ranking
Separated the mixed All score from the People score. Person-only and category-specific Africa/Asia recognition factors and adult-entertainment intent adjustments continue to govern the People leaderboard but no longer reduce a person's score when compared with brands, products, works, groups, and other entity families in All. Cross-entity evidence-reliability rules remain shared. The K-pop Wikipedia/search reliability factor was recalibrated from 80% to 70% so removing the unrelated musician continental factor from groups does not reverse the established requirement that BTS remain outside the All Top 100. Both score paths are persisted for auditability and the People API orders by its dedicated adjusted score.
K-pop coordinated-attention reliability adjustment
Applied a 20% reliability discount to the Wikipedia and search contributions of every K-pop-classified entity in Global Famous, including people and groups. Research documents organized K-pop fandom practices designed to increase streaming, search, and platform visibility, making repeated coordinated activity less representative of distinct-person recognition. News, attributable consumption, structural recognition, geographic evidence, and Trending remain undiscounted. The rule is industry-wide and contains no named-person override.
Living American older-musician scope correction
Restricted the age-derived 30% musician cohort to living U.S. musicians strictly older than 80. Deceased musicians continue to qualify solely through the deceased-musician rule, verified rappers retain 30%, and all other living musicians receive 15%. Country uses recorded entity origin and age uses the recorded full birth date. Exactly one continental factor remains applicable per person.
Musician lifecycle and rapper continental scrutiny
Assigned exactly one continental-recognition rule to each musician: deceased musicians, living musicians older than 70, and verified rappers receive the 30% maximum continuous penalty; other living musicians receive 15%. Age is calculated from the recorded full birth date, and rapper eligibility requires an explicit recorded rapper profession—a hip-hop industry tag alone is insufficient. Actor, athlete, historical, and musician penalties remain mutually exclusive rather than stacked.
Actor worldwide-recognition enforcement
Raised the complete 20-country Africa and 27-country Asia test to a 15% maximum continuous penalty for every verified actor and actress. Multi-category people still receive exactly one continental-recognition factor: the established selection logic chooses the applicable cohort rule and never stacks actor and musician penalties.
Politician and athlete worldwide-recognition enforcement
Raised the complete 20-country Africa and 27-country Asia test to a 15% maximum continuous penalty for every verified politician and athlete. The change is universal within those metadata-derived cohorts and makes a severe two-panel deficit materially affect Global Fame instead of retaining roughly 95% of the pre-test score. Other modern cohorts retain their existing calibrated caps.
Top-100 boundary integrity audit
Barred already-established figures from the low-baseline breakout lane when recognition stock is at least 0.75, or at least 0.60 without an attributable output lane. This prevents renewed attention for established political, coaching, and institutional figures from being mislabeled as fame emergence while preserving genuine new breakouts. The existing continental-recognition rules are unchanged, including their intentional restriction to deceased technology figures.
Reality-personality continental recognition
Applied the complete 20-country Africa and 27-country Asia test with a calibrated 7% maximum continuous penalty to verified reality personalities. Eligibility requires the reality-TV industry plus an explicit socialite, reality-television personality, television-personality, or model profession. The rule is category-derived and contains no named-person exception.
Atrocity-linked political notoriety
Expanded the universal criminal-notoriety-without-output cohort to recorded dictators, war criminals, and genocide perpetrators. Eligibility remains profession-derived and still requires no attributable creative, athletic, screen, music, or audience consumption lane. Qualifying figures receive the existing complete 20-country Africa and 27-country Asia test with its stronger 30% maximum continuous penalty; no named-person rule was added.
Continental cohort classification completeness
Closed three metadata gaps in the complete 20-country Africa and 27-country Asia test. Explicit sport-specific competitor professions now qualify even when an entity's industry is mistagged; screen-qualified film and television directors and screenwriters join actors; and deceased technology figures qualify from recorded tech industry plus deceased status. Each new modern cohort uses the calibrated 7% maximum continuous penalty, while any pre-existing stronger musician or pre-1950 rule remains authoritative.
Influencer continental recognition
Applied the strict 20-country Africa and 27-country Asia recognition test with a calibrated 7% maximum continuous penalty to verified influencers and online creators. Eligibility requires a creator industry or internet-personality classification plus an explicit YouTuber, streamer, influencer, content-creator, TikToker, video-blogger, or vlogger profession. Companies and creator-adjacent celebrities without a recorded online-creator career are excluded. Existing stronger musician scrutiny remains in force for musician-influencers, and complete panels are mandatory in the People Top 100.
Criminal-notoriety concentration control
Applied a stronger 30% maximum continuous Africa/Asia penalty to people whose recorded profession explicitly identifies severe criminal notoriety and who have no attributable creative, athletic, screen, music, or audience consumption lane. The universal metadata-and-evidence rule covers sex offenders, serial or mass murderers, terrorists, drug lords, organized-crime figures, and explicit criminals; it does not target named people or penalize ordinary controversy. Complete 20-country Africa and 27-country Asia panels remain mandatory in the People Top 100.
Actor continental recognition
Applied the strict 20-country Africa and 27-country Asia recognition test with a calibrated 7% maximum continuous penalty to verified actors and actresses. Eligibility requires an explicit actor or actress profession, while the generic profession 'Acting' additionally requires a registered screen industry. Directors, producers, presenters, and acting political officeholders without a performer profession are excluded. Existing stronger musician scrutiny remains in force for musician-actors, and complete panels are mandatory in the People Top 100.
Complete musician continental panels
Expanded the musician Africa/Asia recognition test from six African and nine Asian countries to the complete 20-country Africa and 27-country Asia panels used by strict cohorts. Musician-specific reference, meaningful-recognition threshold, continuous deficit weights, and maximum penalty remain unchanged. Every affected People Top-100 musician must now carry all 20 African and 27 Asian provider observations before publication.
Calibrated athlete continental scrutiny
Calibrated the verified-athlete Africa/Asia adjustment from a 15% to a 7% maximum continuous penalty after complete-panel simulation showed the stronger setting caused excessive cross-category displacement. The weaker-panel deficit may reduce the score by up to 4.667% and the second-panel deficit by up to 2.333%. Athlete eligibility and the mandatory 20-country Africa and 27-country Asia evidence gate are unchanged.
Athlete continental recognition
Applied the strict 20-country Africa and 27-country Asia recognition test with a 15% maximum continuous penalty to every verified athlete. Eligibility requires a person entity, a registered sports industry, and an explicit competitor profession. Teams, competitions, administrators, and coaches without a recorded playing career are excluded. The weaker-panel deficit may reduce the score by up to 10% and the second-panel deficit by up to 5%; complete panels are mandatory in the People Top 100.
Foundational religious figure exemption
Exempted foundational religious prophets from the pre-1950 and political-association Africa/Asia test. Eligibility is metadata-derived: the entity must be a historical figure with an explicit prophet profession. This currently covers Jesus Christ and Muhammad without a named-person allowlist, while modern religious leaders, cleric-politicians, monarchs, and revolutionary figures remain subject to their applicable cohort rules.
Stronger pre-1950 continental scrutiny
Raised the maximum continuous Africa/Asia recognition penalty from 7.5% to 15% for every person born before 1950. The weaker-panel deficit may reduce the score by up to 10% and the second-panel deficit by up to 5%. Politicians, political associates, footballers, and modern musicians retain their existing cohort caps. This category-wide rule strengthens current worldwide-recognition scrutiny without a named-person exception or hard cutoff.
Political-association continental recognition
Applied the strict 20-country Africa and 27-country Asia recognition test to people with reproducible political association evidence. Eligibility covers political types, industries and explicit roles; a structured relationship to a political person; or an associated-with record naming at least two political people. The multiple-name threshold admits documented networks such as Jeffrey Epstein's while rejecting incidental single-person associations. The existing bounded 7.5% maximum proportional influence and complete-panel publication gate apply.
Footballer continental recognition
Applied the strict 20-country Africa and 27-country Asia recognition test to every verified current or former association-football player. Eligibility requires a person entity, the soccer industry, and an explicit football-player or footballer profession; clubs, competitions, administrators, creators, and non-playing coaches are excluded. The same bounded 7.5% maximum proportional influence and 20-observation publication gate apply.
Broad continental recognition panels
Expanded the strict recognition test from six to twenty African countries and from nine to twenty-seven Asian markets, and applied it to every politician as well as every person born before 1950. Africa includes North, West, East, Central, and Southern Africa, including Egypt and Morocco; Asia includes South, East, Southeast, and West Asia. Every affected Top-100 person must have at least twenty provider-measured observations in each panel. Because the panels are much broader and harder, their proportional influence is capped at 7.5% rather than 22.5%. China remains excluded because Google is not a comparable general-search instrument there.
Stronger legacy recognition standard
Raised the pre-1950 Africa/Asia panel target from 0.60 to 0.75 and the meaningful-country floor from 35% to 50% of each subject's own reference. Kept the penalty proportionate at a 22.5% maximum so the standard tests current cross-continent recognition without overpowering the complete Global Fame Score. Modern musicians retain their separately calibrated policy.
Legacy cross-continent recognition
Applied the existing continuous Africa/Asia recognition factor to every person with a recorded birth year before 1950, not only musicians. Any affected person in the flagship Top 100 must have complete comparable panels in both continents. The cutoff is objective and applies equally to living and deceased people.
Durable country fame
Rebalanced Famous IN toward durable country recognition: 45% 2004-present country search, 25% category-neutral passive Recognition Stock after direct country qualification, 15% 28-day country readership, and 15% chart history after at least 28 elapsed days. Young chart runs remain available to Trending IN but cannot dominate Famous merely by straddling a weekly boundary. Added the US Bieber benchmark to the atomic pre-publication gate.
Category-neutral recognition stock
Added a five-percent durable passive-recognition signal based on the number of language Wikipedias carrying a verified Wikidata subject. Existing signals retain ninety-five percent of their prior weight. An exact Wikidata founder relationship may add a small bounded corroboration from only the strongest founded institution. Wealth, net worth, controversy, executive roles, name matching, and hand-picked individual boosts are excluded. The worldwide geography collector now considers prospective industry scores, removing the loop that prevented otherwise-qualified people from receiving the evidence required for a global ordinal.
Responsive worldwide trending
Corrected the shared-anchor scale of weekly search history, prevented a flat media fallback from hiding acceleration measured by another source, admitted TMDB discovery momentum for screen performers, and blended live prominence into Global Trending while keeping acceleration dominant. Durable Famous scores remain separate and unchanged in purpose.
Comparable recent search inside countries
Added a repeatable three-month country-search sweep on the same stable anchor as historical search. Recent values are withheld until at least 100 entities share the comparison scale, and complete sweeps atomically replace partial snapshots so collection order cannot advantage a person.
Comparable historical search inside countries
Withholds anchored historical country-search values until at least 100 entities share the same country comparison scale. Small targeted repair batches remain stored as evidence but cannot give their few measured entities an artificial Famous IN advantage over unswept people.
Smooth musician worldwide-reach adjustment
Replaced the musician Africa/Asia pass-fail tiers with a continuous deficit adjustment. A musician just below the recognition threshold now loses only proportionally, while deep weakness in both continents can still reduce the score by the full 30%.
Durable fame stays above breakout acceleration
Reduced the maximum Famous breakout adjustment from 3.5 to 2.5 score points. Genuine worldwide emergence can still move the ranking, while a recent breakout is less able to overtake a clearly stronger durable fame foundation. Full acceleration remains visible in Trending.
Comparable durable creator fame
Replaced additive YouTube views and subscriber proofs with one bounded creator-durability lane. A top-end curve preserves meaningful separation among very large attributable audiences without increasing the old combined ceiling, billion-view established creators cannot receive a Famous breakout bonus, and publication no longer forces named creators into the Top 100.
Musician-only cross-continental recognition
Restored the v5.21 worldwide ranking foundation and limited the Africa/Asia recognition adjustment to musicians. Creators, actors, athletes, politicians, brands and other entity families are unaffected. Musicians retain their base score when both continent panels are strong, retain 90% when one is weak and 70% when both are weak.
Measured recognition across Africa and Asia
Added a bounded cross-continental recognition factor across worldwide scores. This broad application was retired in v5.23 after it caused unintended cross-category movement.
Durable fame, separate from current attention
Removed the additive current-relevance lane from Famous so a news cycle cannot reverse stronger lasting worldwide evidence. Trending retains the complete current-acceleration measurement. Breakout credit remains available only when the preceding three-year worldwide-search baseline was genuinely low, preventing renewed attention around established legacy figures from being mislabeled as a breakout.
Broad reach over distributed niche intensity
Strengthened the worldwide-reach test across every person and entity. Multilingual concentration now has meaningful score impact, current relevance is capped below enduring fame evidence, and a current-relevance bonus requires both language and geographic breadth rather than allowing strength in one to conceal weakness in the other. Short-lived acceleration remains fully visible in Trending.
Even balance of lasting and current worldwide fame
Raised the long-term worldwide-search foundation from 45% to 50%. Lasting recognition and the combined current Wikipedia, search, media and attributable-audience evidence now receive equal shares for people and groups.
Time-decaying recognition for recent former national leaders
Extended the evidence-qualified national-leader applicability rule to verified former heads of state and government. The normalization remains full for four years after leaving office, fades linearly to zero between years four and eight, and expires completely after eight years.
Comparable evidence for every worldwide people rank
Withheld a definitive Global Fame rank from people whose worldwide evidence is still being collected. Local and Trending evidence remains available, while the worldwide ordinal now requires current multilingual readership, complete applicable signals, at least 20 long-term search observations and geographic coverage across at least 60 tracked countries.
Neutral treatment for countryless historical figures
Stopped treating a missing modern nationality as proof that an ancient or historical figure is foreign to every country. These figures now retain measured local evidence without receiving origin status, origin protection or automatic local fame; modern records with missing metadata remain under the foreign policy.
Current worldwide relevance for brands and products
Extended the existing bounded current-relevance lane to brands, products, companies and organizations. A bonus requires corroboration across live signals plus multilingual and geographic breadth; it is capped at 7.5 points and decays automatically.
A more responsive balance of lasting and current fame
Reduced the long-term worldwide-search share from 60% to 45%, preserving enduring recognition while allowing major current changes in readership, search, media and attributable audience to change the lead. The adjustment applies consistently to all people and groups, never to a named individual.
Evidence-qualified national leaders
Allowed a verified current head of state or head of government to be scored against the local pillars that apply to the role, but only after clearing independent country readership, durable search and news thresholds. Ordinary politicians receive no advantage.
Comparable artists and groups
Applied the same long-term worldwide-search foundation to people and groups so bands do not retain full short-window weight while solo artists retain only 40%, without changing specialist scoring for teams, brands or products.
Fair local evidence applicability
Stopped treating structurally inapplicable local evidence, such as music charts for politicians, as missing entity coverage.
Stable source-specific media evidence
Prevented incompatible news instruments from sharing historical baselines and made source selection compare independently stabilized evidence.
Stronger separation of local and worldwide fame
Separated local and worldwide ranking evidence so changes to a country-origin ranking cannot alter who is measured as famous inside that country.
Comparable worldwide evidence for country-origin rankings
Strengthened country-origin rankings so people are compared only after the Engine has enough worldwide evidence to measure them consistently.
Stronger local qualification
Local fame lists now require stronger evidence that a name is meaningfully recognized or covered in the place, rather than allowing one isolated public trail to qualify on its own.
A stronger view of fame in each place
Improved how the Engine recognizes lasting local interest, preserves meaningful differences among widely known names, and avoids counting repeated chart snapshots as separate periods.
Stronger continuity across data sources
Improved how the Engine maintains a consistent worldwide view when independent data sources refresh at different times or temporarily return incomplete results.
A more responsive living ranking
Improved how the Engine recognizes meaningful changes in worldwide attention, helping established names and genuine breakout figures move as the world changes.
Stronger interpretation of public attention
Improved the Engine's ability to understand what different kinds of public activity mean instead of treating every digital action as identical.
A wider global view
Improved how the Engine recognizes the difference between concentrated attention and fame that reaches people across the world.
Better worldwide and breakout recognition
Expanded the Engine's understanding of where fame lives and strengthened its ability to recognize genuine global breakout names.
The modern Global Fame Engine
Introduced a stronger generation of the Engine built to understand lasting fame, present-day attention, and worldwide recognition together.
What a rank means
A Global Fame rank is the Engine's best current estimate of sustained public recognition. It is not a judgment of talent, character, achievement, approval, or importance. Because public recognition changes, live rankings can change too.